How Social Listening And Machine Learning Are Used To Predict Bitcoin Price Volatility

By Bitcoin Magazine - 2 years ago - Reading Time: 17 minutes

How Social Listening And Machine Learning Are Used To Predict Bitcoin Price Volatility

New data analysis methods can predict bitcoin price movements based on social sentiment. This introduces a new front in the war on FUD.

“Social media” is a gigantic umbrella term that nobody can precisely define. Twitter, Reddit, Facebook, 4Chan, you name it. It doesn't matter if it's a worldwide platform or a tiny forum chat; if it connects users over the internet, it's social media.

On sites like Twitter, users are separated into tiers of opinion authority based on followers and likes. Those that achieve the greatest opinion authority are recognized with the coveted blue checkmark. These blue checkmarks are a badge of authority, which causes an influencer's recognition to continue growing exponentially.

This process has made some social media accounts so powerful that they now have the unique ability to influence bitcoin's price with something as simple as a tweet. Here are some prominent examples.

Given this dynamic, I aim to understand if social media data analytics, more professionally referred to as "social listening" or "emotional data intelligence," can be used to intentionally and effectively influence the public sentiment of bitcoin.

Can emotional data algorithms (EDAs) quantify the exact influence that a celebrity like Elon Musk has over the price of bitcoin?

To accomplish this, I will apply insights from deep research of emotional data intelligence companies like Stockpulse, actionable insight companies like S2 Research, and multiple collegiate studies regarding the relationship between social media influence and bitcoin price volatility.

Ultimately, this study aims to illustrate a new important framework for thinking of machine learning concepts like social listening and emotional data intelligence. I will also open a discussion challenging the ethics of new concepts like “emotional data algorithms” and "digital language processing" (also known as "sentiment analysis"), as cited by Stockpulse CEO Dr. Stefan Naan.

Ultimately, new technological phenomena have catapulted the study of social listening into a powerful tool for predicting public opinion and bitcoin price volatility.

Our Digital Role Models

I will be examining how "social listening," celebrity influence and the advent of EDAs can all be used to predict and affect bitcoin price volatility. Before starting this research, my initial personal dogma leaned on the idea that social media is just unpredictable chaos. I thought, "Nobody can keep track of it all." Like most people, I believed that nobody could control and manipulate social media.

Why bother worrying about it? Right?

The evidence that I found proved otherwise. Rapid developments in machine learning, social media data collection, and predictive algorithms have proven that social media chaos can be organized into actionable insights.

It's not uncommon for people to look to popular icons as sources of immense inspiration, treating them as our own personal digital role models.

This isn't necessarily a bad thing. As with most things, it's complicated. These social media figures guide us in ways that people in our real lives sometimes can't. It's fair to assume that they primarily give people well-intentioned advice. However, there are also a great many trusted influencers that have sold out their audiences to pump assets and grow their portfolios.

You cannot believe everything you read.

What Is Social Listening?

Social listening is defined as "social media measurement and social media analytics; a way of computing popularity of a brand or company by extracting information from all social media channels, such as blogs, wikis, news sites, micro-blogs such as Twitter, social networking sites, video/photo sharing websites, forums, and message boards," by Wikipedia.

Take the study of movie fandoms like “Star Wars” as a good example of powerful social listening. “Star Wars” has a notoriously passionate legion of fans. You've seen them; they are a mighty unit that is occasionally hostile and extremely influential. The opinion of that specific group of viewers is critical to the success of Disney movie producers and the longevity of their careers. The views of these fans can set the groundwork for a film's long-term legacy. As such, predicting the behavior of fandoms is an increasingly sought-after commodity in the film industry.

Movie fandoms are just one small example of how social listening has become a high-value area of study. As newer, more disruptive technologies like Bitcoin become increasingly powerful, those that influence sentiment about Bitcoin also become increasingly powerful.

How Does Social Listening Affect Bitcoin Prices?

Social listening becomes exponentially more profound when you tie in its relationship to the world's largest financial revolution, Bitcoin.

According to Naan, writing for Nasdaq, "By monitoring and analyzing data from social media sources — especially concerning communication about stocks — it's now possible to connect the dots between sentiment and market movements."

Even the largest stock exchange in the world has perked its ears up to the weight that social media data has on market prices.

The results suggest that our quantitative understanding of how influencers can use emotional data to sway public perception of assets like bitcoin is much more advanced than most people realize. The surface level correlations are quite apparent after compiling examples of celebrity/influencer tweets causing market fluctuations. However, the hyper-specified nature of these data-driven insights and how they are used is of larger importance.

André Kostolany, a stock market investor who became one of the most successful investors of the 20th century, has said that "facts only account for 10% of the reactions on the stock market; everything else is psychology."

Wealthy and influential public figures worldwide are now consciously aware that they have a strong ability to manipulate public sentiment. Many celebrities have learned how to instantly incite revolution or pump their favorite coins and stocks with just 280 characters. In the era of the Bitcoin revolution, influencers ranging from celebrities to billionaires to politicians have all begun to flex their social clout to shift Bitcoin sentiment in their favor.

This creates a virtual tug-of-war for control of bitcoin's positioning in the public eye.

How Is Social Listening Calculated?

Now that you know what social listening is in a general sense, you're probably wondering how the heck people can make sense of all that data. I found a company that serves as a strong example of social listening in action: Lunar Crush.

Lunar Crush is a social listening and analytics company that allows you to track the social mentions and public perception of Bitcoin in rea -time. It creates detailed rankings of digital coins based on metrics like “social dominance," “sentiment analysis” and “influencer rankings.” All of these metrics are used to create data-driven insights for investors.

Its site explains that "Social Dominance calculates the ‘share of voice’ across all social media data. This is similar to Market Dominance; however, instead of dividing a coin's market cap by the entire cryptocurrency market, we divide a coin's social volume by the entire cryptocurrency market's social volume."

For example:

Bitcoin social volume: 1,000,000 mentions

Entire cryptocurrency market social volume: 10,000,000 mentions

1,000,000/10,000,000 = 10%

Bitcoin social dominance in this example would be 10%

In other words, data-driven insights, powered by social listening metrics, give bitcoin analysts the ability to make decisions based almost solely on the number of times “Bitcoin” has been mentioned across much of social media.

Emotional data intelligence doesn't only calculate the frequency of “Bitcoin” mentions across social media. It is now possible to predict bitcoin prices using what's referred to as "sentiment analysis."

An EDA can crawl through archives of every tweet, post and comment across all of social media, analyzing each for keywords that indicate "bullish" or "bearish" sentiment of a chosen digital asset such as bitcoin. This data is then translated into human digestible market insights based on overall public opinion.

A New Gold Rush For Data Analytics

In this digital age, public sentiment is everything. With it, nothing can fail; against it, nothing can succeed. Whoever molds public sentiment goes deeper than they who enact statutes or pronounce judicial decisions. Social media influencers, celebrities and public figures are the weavers of our shared narrative. Thus, it is imperative to understand the magnitude of their influence.

These claims of social listening's newfound importance are backed up when German emotional data intelligence company Stockpulse notes, "While the technologies themselves may not be new, the interest from exchanges and banks is. To date, the tech companies leading the software development have largely served hedge funds, seeking a competitive market edge. Major exchanges have now jumped on board."

The technology to predict and plan for shifts in public sentiment already exists, made readily available using predictive algorithms and data collection. Major institutions and titans of industry are already instituting these quantitative analysis protocols. That being said, most people are still largely unaware that these technologies even exist.

Emotional Data Intelligence Used To Develop Sentiment Analysis

Stockpulse is a machine learning/AI tech company that has developed data analytics software and algorithms to quantitate social media data into actionable investment insights. It works primarily with what it describes as "Digital Language Processing" (DLP), also referred to as "sentiment analysis."

For Nasdaq, Naan wrote that "Stockpulse collects and analyzes data from social media sources around the clock in German, English, and Chinese… We have historical data from alternative sources that date back to 2011. Our web crawlers are continuously scanning thousands of different Internet sources for relevant financial topics and communication, collecting several million tweets, chat messages, message board posts, news articles, and comments to news articles each day.”

The claims about web crawlers collecting massive stores of all social media interactions are purely factual. Many reputable sources echo these claims.

"Emotional Data Intelligence" can be used to create quantifiable rankings of public figures based on the potential to influence markets.

According to Naan, in the article written for the Nasdaq, "Some market participants potentially have a higher impact on the movement of prices than others."

Stockpulse has developed a curated list of verified social media users (including Twitter accounts of CEOs of listed companies, influential politicians, journalists, analysts, and news agencies) and ranked their ability to influence markets using predictive algorithms, data collection and machine learning.

The applications of such technology go far beyond curated lists.

"In addition to detecting the expert network of social media users for a single stock or industry, an alert system for specific Twitter accounts could be highly relevant,” Naan wrote for Nasdaq. “As soon as a Twitter user identified as a relevant and credible source for a stock posts anything about that company, trading surveillance wants to know about it. Having this information as quickly as possible is key."

In regards to Bitcoin, similar predictive algorithms can rank public figures based on their ability to influence bitcoin's price. You can now find data-driven rankings boards for Twitter users with "potential to influence Bitcoin." The attention economy has officially been quantified.

Take Musk, for example. It's no secret that he would currently top the list of public figures with the ability to influence bitcoin. As such, It would be ideal to program an algorithm to place a buy or sell order precisely when Musk tweets something that contains the keyword "Bitcoin." If the algorithm indicates bullish sentiment, you buy, and if it indicates bearish sentiment, you sell.

A ranked list of influencers would not only visualize which public figures hold the greatest sway. Investors would highly monitor it to alert users of precisely the moment that one of these influencers posts something that would affect the bitcoin price.

How Does Emotional Data Intelligence Impact Bitcoin Price?

To answer this question fairly, we have to take steps back and look at how bitcoin interacts with social media differently than do traditional stocks.

According to Korean scientists Sejung Park and Han Woo Park, in their academic study called "Relations Between Reputation And Social Media Communication In Cryptocurrency Markets," "Bitcoin posts were more frequent when the coins' prices were high and less frequent when the prices were low."

In other words, Bitcoin influencers attempt to time their marketing messages to the highs and lows of the market, allowing public figures to maximize their potential influence over market buys.

This isn't a shocking, new phenomenon by itself, but now EDAs give marketers the ability to use this same strategy using precise social dominance metrics instead of loosely timing price volatility.

Influencers and brands have become attached to this new, more powerful form of marketing which entices new investors using excitement, hype and memes in concurrence with timed postings. As long as they remember to say "this is not financial advice," there's not any liability.

This is not financial advice. See? :)

Gourang Aggarwal, a researcher for NIIT University in India, calculated tangible results for this concept in his paper "Understanding Social Factors Affecting The Bitcoin Market." Aggarwal wrote, "Opinions of social media influencers having a relationship with Bitcoin-industry (negative) has a correlation value -0.0631 which shows that a negative news story from a big personality or celebrity might drop the Bitcoin price."

He continued, "Media messages that relate to the keywords ‘Bitcoin bans’ (negative) (the 6th attribute) correlate with a value of -0.1329 and can also lead towards a drop in the price of Bitcoin.”

In layman's terms, it is statistically true that any news about negative buzzwords (aka, FUD) like "Bitcoin bans," "ChinaBan," or any other negative sentiment from celebrities generally results in a price drop. Duh. But it gets more insightful when you realize that spreading these messages is a deliberate attempt to use EDAs to sway public opinion of new emerging technologies like blockchains and Bitcoin.

Many opponents of Bitcoin aim to drain prices by spreading negative sentiments about the technology. They frequently use FUD (fear-based buzzwords) to scare away potential defectors of the legacy system. EDAs allow those who spread FUD to do so with drastically increased hyper-specificity. In other words, they can use emotional data intelligence to target FUD towards those who are most likely to believe it.

Looking at you, governments, central banks and hedge funds ;)

Using EDAs To Track The Epidemic-Like Spread Of Investment Ideas

The extent to which our data are used and commoditized is still a mystery to most people. According to Naan, "People interacting on social media generate emotional data by expressing their emotions and opinions via tweets, forum posts, and blogs. They also consume it, and in the process are influenced by the sentiments, feelings, and opinions expressed by others."

Social media data have moved far beyond location tracking and TikTok recommendations. Our applications now can track our emotional habits and influence our opinions with "hyper-targeted" conversion placements.

For more about the dangers of this concept, check out the Netflix documentary “The Social Dilemma.”

Concerning Bitcoin, Ross Christopher Phillips of University College London has said that "This work demonstrates how marketers can apply epidemic detection techniques to social media data to predict Bitcoin prices and provides some empirical evidence that ‘bubbles’ mirror the social epidemic-like spread of an investment idea."

This butterfly effect concept creates a potential intersection between EDAs and bitcoin volatility prediction.

Similarly, Mehrnoosh Mirtaheri, USC Information Sciences Institute, wrote about using social listening metrics to track "pump and dump schemes" in his paper "Identifying And Analyzing Cryptocurrency Manipulations In Social Media."

He wrote, "Specifically, given financial and Twitter data pertaining to a particular coin, our method can detect, with reasonable accuracy, whether there is an unfolding pump and dump scheme, and whether the resulting pump operation will succeed in terms of meeting the anticipated price targets."

Algorithms and machine learning have developed a method capable of predicting the occurrence of "pump and dump schemes" for specific digital tokens. It can also reasonably predict their chances of success.

Phillips echoed this confident analysis, writing that "The authors found the relationship between price and Twitter submissions acts as an amplification mechanism; a positive feedback loop is identified whereby firstly price increases cause search volume to increase, which in turn causes mentions on Twitter submissions to increase, with this, in turn, causing a further price increase."

In other words, there is a tangible sequence of events that occurs whenever bitcoin's price fluctuates dramatically.

When a bullish event causes a price rise, an epidemic-like spread of hype and attention circulates, which causes the bitcoin price to rise even more. The same is true when applied to a bearish scenario.

“The Starstruck Effect” Epidemic-Like Spreading Of Social Media Investment Ideas

Did you ever hear from someone that Musk wasn't tweeting to manipulate prices but rather that he was "experimenting"?

This is what he was experimenting on.

Universities worldwide have conducted several highly-relevant studies to gauge whether celebrity influence over market volatility can be tracked and predicted. The two key points of research methodology used include comparing social media user data to stock market fluctuations.

In simple terms, if someone posts about Bitcoin on social media, in particular on Twitter and Reddit, algorithms can track if that post has in any way caused bitcoin’s price to trend upwards or downwards. People who are good with numbers can follow this and make investment decisions based on those results.

Phillips wrote that, "Data favors a more epidemic-like definition, describing a price fluctuation as occurring by psychological contagion, where the news of price increases spurs investors' enthusiasm which spreads contagiously and brings in a larger group of investors, drawn in by envy and excitement about the previous price rises.”

Another framework for understanding this: As bullish news spreads across social media in an epidemic-like fashion, the bitcoin price will rise epidemically, too. Similarly, bitcoin's price will fall accordingly if FUD or bearish news is spread in an epidemic-like manner.

Another way to look at this: Bitcoin fluctuates according to the opinion of the masses. Decentralization in all of its glory.

Popular influencers like Musk have realized that their social media accounts are an increasingly powerful financial tool. When celebrities post about exciting "investment opportunities," people begin to flock to the asset because they already hold strong pre-existing positive associations with that celebrity. Furthermore, the online fandoms of that same celebrity start to share and amplify the investment idea until an epidemic-like spread is created. This is also why new phenomena like the price of AMC and GameStop shares are so resilient.

The Case Of Elon Musk And Bitcoin

An example of this in the Bitcoin arena would be Musk boasting about his love of Bitcoin and tweeting about Tesla's intentions to invest in the asset. Long-term holders of bitcoin profit greatly off of Musk hype due to the influx of new buyers it generates. Afterward, those same bitcoin "HODLers" start to engage in an epidemic-like spread of posts of their own, bragging about their gains and spreading memes. This contagion-style hype spread causes bitcoin prices to pump dramatically. This also inevitably causes less experienced investors to buy during bad price points due to FOMO (fear of missing out).

When Musk tweeted bullish Bitcoin content, such as the news of Tesla buying in or the occasional positive meme, the bitcoin price soared to the moon. When he tweeted "bearish" sentiments, such as environmental concerns for Bitcoin mining, the prices tanked accordingly.

Corporate Propaganda Or Revolutionary Rhetoric?

What do you think about algorithms being used to track and influence our emotions?

Bitcoin is a revolutionary technology that offers to solve some of the world's greatest injustices. As such, it seems that it is ethical to use any means necessary to spread the good word. However, governments and private institutions can also use EDAs to influence public opinion for an innumerable amount of other subjects. It seems to me that this realization does indeed warrant a thorough ethical examination.

Philosophically speaking, revolutionary rhetoric and propaganda are closely related but also fundamentally different. Both are linked to the epidemic-like spread of messages, as cited earlier.

However, the ideas and intentions behind them couldn't be farther apart.

Revolution, in its purest intended definition, stems from the concept of defending basic human rights. It is the idea that if governments don't treat their citizens fairly then those citizens will, and should, band together to demand a system change. This unity is exclusively achieved through the epidemic spread of revolutionary rhetoric.

It's a fundamental function of the "American dream." Generally speaking, revolution stems from good intentions. It's about freedom and equality above all else. Sometimes, people go off the rails and break things, set fires, all that crap. That garbage is unfortunate, but it is also an inevitable human behavior.

Regardless, revolutions almost always stem from classic American ideals of freedom and fair competition.

In contrast, propaganda stems from ambitions of societal control. Those in power use propaganda to get people to agree with certain philosophies and ideas without explaining the fundamental concepts behind them.

In other words, they expect you to "take their word for it."

Widespread messages of these magnitudes, if followed blindly, if left unquestioned and unsourced, are much more akin to propaganda. In other words, Bitcoiners can use EDAs and social listening metrics to spread a positive new alternative, but legacy institutions can also use these algorithms to spread propaganda and brainwash the public.

Ask yourself about the long-term implications of this relationship between EDAs and market prices. What happens when someone controls the messages and, simultaneously, the messages control the money?

I will be outlining that the power of social media influencers, if ignored and unquestioned, is potentially a massive hindrance to the concept of "fair competition."

There are good arguments for seeing this technology as bad for society rather than good.

Naan wrote that, "We don't know anything about the person, but there are metrics with probability scores where you can identify if a person is more important than another person.”

It seems that algorithms determining "human importance" raise potential ethical concerns.

And Naan noted another potentially harmful development: "Trading surveillance teams can monitor any rumors or posts about relevant events in real-time in social media and get instant notifications if certain companies or events suddenly move into focus."

These power players get real-time access to all the premium data-driven insights. The competition gap between "average joe" investors and wealthy/institutional investors has increased exponentially with the implementation of EDAs.

On the one hand, politicians, central banks, hedge funds, and the media often use their wealth, power, and influence to make themselves richer and consumers poorer. They often leave investors behind in a vicious cycle of giving one and taking two.

On the other hand, Bitcoin influencers (the reputable ones) use their wealth, power and influence to promote a new technology with a strong potential to increase financial inclusivity worldwide.

The use cases for blockchain technology are next to limitless, potentially fixing countless systemic problems constantly present throughout our lives. Surely this seems to justify the use of EDAs to further the mass adoption of Bitcoin? At the very least, we know that major financial institutions are already using EDAs; Bitcoiners may need to use these technologies if only to ensure a fair fight.

Defending Emotional Data Algorithms

It's important to note that the root of ethical issues is not about technology; it's about the people who control it and use it to manipulate markets or bitcoin prices. It's those with ulterior motives who govern the technology which need to be further examined.

Henry Regler has explained that "With the growth of social media's presence in financial markets, it is perhaps only a matter of time before regulators themselves employ Sentiment Analysis and other tools in their market surveillance. Some legal experts said they wouldn't be surprised if they already have."

There are still many essential questions left unexplored, even after all of the initial technical data research. Is it ethical to track and measure such things as users' social media sentiment? Is public opinion supposed to be controlled and predicted? What are the long-term consequences of leaving these complex algorithms exclusive to big corporations, governments, and hedge funds?

These are things that might sound alarmist upon initial reading. The point is not to assume that everything about emotional data intelligence and celebrity influence must be manipulative or oppressive. Technology doesn't automatically function in the interest of malice, as “Terminator” or “War Games” would illustrate.

What if all of these new powerful technologies and algorithms could be used by rehab facilities to help people with addiction? What if scientists used them to develop mental health research in congruence with social media data? What if they could reduce the negative psychological impacts of frequent social media usage in young bodies?

There seem to be many more unquestionably ethical applications for this technology that don’t just entail making rich investors richer. Decide for yourself how all of these narratives and fintech developments fit into your worldview.

These are my opinions based on a mountain of reading and academic research. I have not been paid to write or research this topic

There are strong reasons to believe that controlled social influence can be harmful and manipulative. There is also strong evidence of strategic social influencing fostering a better widespread understanding of disruptive technologies like Bitcoin and its potential to increase social good.

Further implementation of emotional data intelligence and EDAs could lead to a deep centralized control of public opinion and social media sentiment. EDAs could also accelerate the widespread adoption of Bitcoin, which some see as the largest techno-sociological revolution in human history.

Who’s to decide?

Unsurprisingly, the fate of free will seems to remain in the hands of wealthy influencers and predictive algorithms. Now, it's off to the races.

This is a guest post by Cameron Palmer. Opinions expressed are entirely their own and do not necessarily reflect those of BTC Inc or Bitcoin Magazine.

Original source: Bitcoin Magazine